Meteorologisk institutt met.no

SPARSPAR : :

Rasmus E. Benestad, Yvan Orsolini, Ina T. Kindem, Rasmus E. Benestad, Yvan Orsolini, Ina T. Kindem, Arne MelsomArne Melsom

Seasonal Predictability over the Arctic Region – exploring the role of boundary conditions

Norwegian Norwegian Institute for Institute for Air ResearchAir Research

Motivation

Motivation

• SPAR

Motivation

Motivation

• SPAR

Motivation

• SPAR

Courtesy of ClimateExplorer, Geert Jan van Oldenborgh, KNMI

Ensemble mean correlation

http://climexp.knmi.nl/forecast_verification.cgi

• SPAR

Courtesy of ClimateExplorer, Geert Jan van Oldenborgh, KNMI

Brier Skill Score w.r.t. climatology

http://climexp.knmi.nl/forecast_verification.cgi

Seasonal predictability 'high' in the Tropics (ENSO). Notoriously low in higher-latitudes

Forecast models Initial conditions in the Tropics Sea-ice, snow, SST, role of stratosphere not well

represented

Climate change & media pressure International Polar Year

New observations

Fundamental science How does predictability vary with latitude?

Objectives

Identify signals in the Arctic. Examining predictability in northern

Europe

• SPAR

Means

Numerical experiments ECMWF IFS model

Analysis Model data Observations (IPY?)

• SPAR

Seasonal forecast models apply prescribed sea ice, alternatively, initial fields are strongly relaxed towards climatology.

…but sea ice undergo interannual fluctuations, e.g. Deser et al., 2000

SPAR WP1: Sea-ice

Is the sea-ice in seasonal forecast models a limiting factor for the quality of these forecasts at high latitudes?

WP1.1: Winter sea ice patterns and summer ice extent extremes

WP1.2: A summertime “blue Arctic”

WP1.3: Limits of seasonal predictions with “perfect” sea ice data

N.H. September ice extent,CCSM/SRES-A1B

(Holland et al., GRL ’06)

NOAA oceanexplorer

• SPAR

Eurasian snow cover has been linked to the forcing of large-scale NH teleconnections. Extensive fall snow cover has been linked to a negative NAO phase in the following winter.

(e.g. Cohen et al., 2004; Saito et al., 2001)

SPAR WP2: The role of snow cover

Recent 20-year ensemble simulations with an AGCM forced by global satellite snow cover observations indicates a better representation of the year-to-year variability in the Icelandic and Aleutian Lows, and their out-of-phase coupling, compared to control simulations.(Orsolini, Kvamstø and Sorteberg, to be submitted, 2007)

We aim at improving the use of snow variables in initialisation of seasonal forecasting, and assess the impact of snow cover inter-annual variability onto the NH high latitude circulation.

• SPAR

Impact on 500mb wind speed : note

Negative NAO-like signal

SPAR WP3: The stratosphereSSW composite

We aim at examining three issues:

1: SSWs and Arctic storm track changes: While stratospheric extreme events influence mid-latitude storm tracks (see figure to the left), do they specifically influence the path of cyclones into the Arctic ?

2: Predictability of the spring onset: Can the occurrence of the stratospheric final warming lead to improved predictability in the troposphere (like SSWs in mid winter) ?

3: Predictability of SSWs: we will carry out ensemble medium-time-scale forecasts during the IPY to better understand precursors of SSWs, and synoptic conditions during their downward influence.

Simulations with Arpege GCM (T106-60lev)

Kindem, Orsolini and Kwamstø, to be submitted, 2007.

SST patterns are associated withthe large-scale atmospheric circulation,not only in the tropical region (El Niño), but also at mid-latitudes(e.g. the NAO/SST tri-pole pattern.)

(Rodwell et al., 1999; Benestad & Melsom, 2002)

SPAR WP4: SST

Examine how seasonal forecasts are affected by perturbations of anomalies on a large scale, withan emphasis on SST and sea ice consistency

WP4.1: The NAO-tripolar SSTA pattern

WP4.2: Arctic SSTA patterns

WP4.3: Simulations with different polar SSTAs

N. Norway

Svalbardarchipelago

Barents Sea75N

30E

SST from IR:

+ Data coverage+ Front resolved- Resolution- Coastal data

SST from IR:

SST from microwave:

Winter data

• SPAR

StratosphereSST

Sea-iceSnow

cover

Picture from ozonewatch.gsfc.nasa.gov

Meteorologisk institutt met.no

Meteorologisk institutt met.no

Extra slides

• SPAR

Figure copyright:

David W. J. Thompson

The troposphere responds to changes in the stratospheric state.

(e.g. Wallace and Thompson, 1998, Baldwin and Dunkerton, 2001)

SPAR WP3: The stratosphere

NH winter While numerical simulations with troposphere-stratosphere AGCM indicates downward influence of extreme stratospheric events (such strong vortex episodes, or stratospheric sudden warmings), there is mild impact on winter seasonal forecasts. Rather the influence is on the medium time scale.

We aim at improving practical use of the S-T coupling in sub-seasonal and seasonal forecasts

• SPAR

SST data in SPAR:

• SST from existing seasonal forecasts(control experiment)

• SST from IR measurements(sensivity to ocean model trends & errors)

• SST from microwave data(spatio-temporal data coverage and instrumentation)

Life-cycle of stratospheric

sudden warmings:

geopotential composites

Winter

Results•Precursory High anomalies over northern Europe, akin to Scandinavian blocking

•Lingering negative NAO anomalies near surface (as shown by Baldwin et al., or Limpasuvan et al)

ONSET

GROWTH

MATURE

DECLINE

DECAY

100mb 1000mb

Nao negative

30mb 500mb 1000mb

Precursory Blocking

Lingering NAO-

Simulations with Arpege GCM (T106-60lev)

Kindem, Orsolini and Kwamstø, to be submitted, 2007.

• SPAR

SST data in SPAR:

• SST from existing seasonal forecasts(control experiment)

• SST from IR measurements(sensivity to ocean model trends & errors)

• SST from microwave data(spatio-temporal data coverage and instrumentation)

• SPAR

SST from IR measurements(traditional method)

vs.SST from microwave data(experimental method);

winter data

N. Norway

Svalbardarchipelago

Barents Sea75N

30EC

+ data coverage+ front resolved- resolution- coastal data

K

IR measurements vs. microwave data;summer data

• SPAR

C K

• similar features• microwave data are warmer in the southern Barents Sea

IR measurements vs. microwave data;summer data trends over an 8 day period

• SPAR

• obvious differences, e.g. in the central Barents Sea• ‘noisy’ trends in the high-resolution IR measurements

Collaboration with EC-Earth Complimentary studies

Use experimental set-up for comparisons Share data Use EC-Earth data as input? Different focus but overlapping interests

Share experience Modelling Data processing

• SPAR

• SPAR

Time schedule:July 2007: Hire postdoc.

Start setting up modelAnalysing input data

2008-2009: experimental runs

2010: analysis and reporting.

• SPAR

• SPAR

• SPAR